Marriage and Cohabitation in western Germany and France

Transcription

1 Marriage and Cohabitation in western Germany and France Dissertation zur Erlangung des akademischen Grades doctor rerum politicarum (Dr.rer.pol.) der Wirtschafts- und Sozialwissenschaftlichen Fakultät der Universität Rostock vorgelegt von Katja Köppen Rostock, Datum der Verteidigung: Erstgutachter: Prof. Dr. Michaela Kreyenfeld 12 Zweitgutachter: Prof. Dr. Heike Trappe 1 1 Universität Rostock, Institut für Soziologie und Demographie 2 Max-Planck-Institut für demografische Forschung, Rostock urn:nbn:de:gbv:28-diss ]

2

3 Acknowledgements This thesis was mainly written during my time at the Max Planck Institute for Demographic Research in Rostock. First of all, I would particularly like to thank my supervisor Michaela Kreyenfeld for her comprehensive support and excellent mentoring through all the years. I greatly benefited from her knowledge and ongoing advice in the field of family sociology and demography. I am grateful for her patience and for putting trust in me. My gratitude goes to Jan Hoem and Joshua Goldstein who gave me the opportunity to stay at the institute and work in the friendly atmosphere of the East Wing. I am particularly grateful to all the former members of the Laboratory of Contemporary European Fertility and Family Dynamics for their support and advice. My appreciation goes to Gunnar Andersson, Gerda Neyer, Hill Kulu, Rainer Walke, Karin Tesching, David Alich, Cordula Zabel, Esther Geisler and Dorothea Rieck. In addition I would like to thank my local supervisor Laurent Toulemon for two busy and inspiring months in Paris at INED. In March 2010 I started working at the University of Rostock. I am especially grateful to Heike Trappe for her support and for giving me the possibility to finish my dissertation. Without the support of my family and friends during all the years of studying, this work would not have been possible. In particular I would like to thank Uta Ziegler, Annett Fleischer, Elke Loichinger, Sabine Schnabel and Anne Hornung. Thanks especially to my parents for their ongoing support and trust. My love goes to my own little family, Uwe and Charlotte, who always showed me that there is a life next and after the dissertation. i

4 ii ACKNOWLEDGEMENTS

5 Contents Acknowledgements i 1 Introduction 1 2 Demographic developments 11 3 Theoretical framework Introduction The economic independence theory Denial or delay of marriage? Changes in women s bargaining position Empirical evidences Summary Ideational theory Second demographic transition Individualization Cultural differences The institutional perspective Summary Contextual framework Introduction Meaning of marriage & cohabitation Family Policies History Child care iii

6 iv CONTENTS Leave regulations Taxation Regulatory framework Summary Educational systems and labor markets Educational attainment and educational enrollment The situation of men and women in the labor market Summary Summary and research hypotheses Changes over time The impact of education Educational level Educational enrollment The impact of employment The impact of children Data and methods The concept of the life course Methods Data The German Family Survey The French Study of Family History Events under study Sample selection Covariates Empirical findings Introduction Descriptive analysis First union formation in western Germany Cumulative incidence curves of first union formation Event history analysis of first union formation Transition to subsequent marriage

7 CONTENTS v First birth and subsequent marriage formation First union formation in France Cumulative incidence curves of first union formation Event history analysis of first union formation Transition to subsequent marriage First birth and subsequent marriage formation First union: Summary of the results Union formation over time The effect of women s education on first union formation The effect of women s employment on first union formation The effect of pregnancy and motherhood on first union formation Personal background characteristics and first union formation Conclusion Introduction Substantial research findings Critical reflections Appendix Appendix A Appendix B Appendix C 269 Bibliography 275

8 vi CONTENTS

9 List of Figures 2.1 Crude marriage rate, Total female first marriage rate, Sum, by five-year age-group, of female first marriage rates (age in completed years) in western Germany, Sum, by five-year age-group, of female first marriage rates (age in completed years) in France, Proportion of non-marital births, selected years Distribution of cohabiting and married couples with children in western Germany, selected years Distribution of cohabiting and married couples with children in France, selected years Labor force participation rates of western German women, selected years Labor force participation rates of western German men, selected years Unemployment rates 1 of western German women and men, selected years Unemployment rates of western German women and men, age 15-24, selected years Labor force participation rates of French women, selected years Labor force participation rates of French and western German mothers by number of children under age 25 and age of youngest child (in per cent), Labor force participation rates of French men, selected years Unemployment rates of French women and men, selected years Unemployment rates of French women and men, under age 25, selected years Median ages for certain life events, western German women, selected birth cohorts (n= ), survival time data analysis 134 vii

10 viii LIST OF FIGURES 7.2 Median ages for certain life events, French women, selected birth cohorts (n= ), survival time data analysis Cumulative percent ever entering first marriage within a first union by selected ages and birth cohort, western German women (n=2.964), frequency tables analysis Cumulative percent ever entering first marriage within a first union by selected ages and birth cohort, French women (n= ), frequency tables analysis Kaplan Meier estimation for the transition to first union by birth year of the women, western Germany (n=2.964) Kaplan Meier estimation for the transition to first union by birth year of the women, France (n= ) Proportion of first unions that started as cohabitation across calendar year of union formation in western Germany 1 (n=2.398) and France (n= ), frequency tables analysis Kaplan Meier estimation for the transition to subsequent marriage by birth year of the women, western Germany (n=1,189) Kaplan Meier estimation for the transition to subsequent marriage by birth year of the women, France (n=64,049) Outcome of western German women s first union after five years. Percentage distribution by mode of entry (cohabitation or marriage) and by outcome (n=2.406), frequency tables analysis Outcome of French women s first union after five years. Percentage distribution by mode of entry (cohabitation or marriage) and by outcome (n= ), frequency tables analysis Union status at first birth by calendar year of birth, only first unions, western German mothers (n=1.779) and French mothers (n=90.539), frequency tables analysis Cumulative incidence of cohabitation and direct marriage for grouped birth cohorts, western Germany (n=2.964) Cumulative incidence of cohabitation and direct marriage for first school graduation, western German women (n=2,920 1 ) Cumulative incidence of cohabitation and direct marriage for women with no degree or Hauptschul degree as first school graduation by grouped birth year of the woman, western German women (n=1,029) Cumulative incidence of cohabitation and direct marriage for women with Realschul degree as first school graduation by grouped birth year of the woman, western German women (n=1,185)

11 LIST OF FIGURES ix 7.17 Cumulative incidence of cohabitation and direct marriage for women with Abitur or Fachhochschulreife as first school graduation by grouped birth year of the woman, western German women (n=706) Piecewise-constant baseline intensity for transition to first union, western German women Piecewise-constant baseline intensity for transition to first cohabitation by birth cohort, western German women Piecewise-constant baseline intensity for transition to first direct marriage by birth cohort, western German women First union formation intensities by type of union and calendar time, western German women First union formation intensities by type of union and calendar time, western German women Relative risk of entering cohabitation by level of education and age of the woman, western German women Relative risk of entering direct marriage by level of education and age of the woman, western German women Relative risk of entering cohabitation by first school graduation and age of the woman, western German women Relative risk of entering direct marriage by first school graduation and age of the woman, western German women Relative risk of entering cohabitation by level of education and birth cohort of the woman, western German women Relative risk of entering direct marriage by level of education and birth cohort of the woman, western German women Relative risk of entering direct marriage by calendar time over Model 1 and Model 5, western German women Piecewise-constant baseline intensity for transition to subsequent marriage, western German women Subsequent marriage formation intensities by age of the woman, western German women Subsequent marriage formation intensities by age of the woman and birth cohort, western German women Subsequent marriage formation intensities by calendar time, western German women Transition to subsequent marriage: effect of first conception and subsequent childbirth, western German women Transition to first conception: effect of a first marriage for cohabiting women, western German women Transition to subsequent marriage: interaction between pregnancy motherhood status and calendar year of union formation, western German women

12 x LIST OF FIGURES 7.37 Transition to first conception: interaction between marriage duration and calendar year of union formation, western German women Cumulative incidence of cohabitation and direct marriage for grouped birth cohorts, France (n= ) Cumulative incidence of cohabitation and direct marriage for highest level of education measured at interview, French women (n=122,606 1 ) Cumulative incidence of cohabitation and direct marriage for women with no or low level of education by grouped birth year of the woman, French women (n=37,867) Cumulative incidence of cohabitation and direct marriage for women with medium level of education by grouped birth year of the woman, French women (n=55,730) Cumulative incidence of cohabitation and direct marriage for women with high level of education by grouped birth year of the woman, French women (n=29,009) Piecewise-constant baseline intensity for transition to first union, French women Piecewise-constant baseline intensity for transition to first cohabitation by birth cohort, French women Piecewise-constant baseline intensity for transition to first direct marriage by birth cohort, French women First union formation intensities by calendar time, French women First union formation intensities by type of union and calendar time, French women Relative risk of entering direct marriage by calendar time over Model 1 and Model 2, French women Relative risk of entering cohabitation by calendar time over Model 1 and Model 2, French women Relative risk of entering cohabitation by level of education and age of the woman, French women Relative risk of entering direct marriage by level of education and age of the woman, French women Relative risk of entering cohabitation by level of education and birth cohort of the woman, French women Relative risk of entering direct marriage by level of education and birth cohort of the woman, French women Piecewise-constant baseline intensity for transition to subsequent marriage, French women Subsequent marriage formation intensities by age of the woman, French women

13 LIST OF FIGURES xi 7.56 Subsequent marriage formation intensities by age of the woman and birth cohort, French women Subsequent marriage formation intensities by calendar time, French women Transition to subsequent marriage: effect of first conception and subsequent childbirth, French women Transition to first conception: effect of a first marriage for cohabiting women, French women Transition to subsequent marriage: interaction between pregnancy motherhood status and calendar year of union formation, French women Transition to first conception: interaction between marriage duration and calendar year of union formation, French women 222

14 xii LIST OF FIGURES

15 List of Tables 2.1 Development of selected demographic indicators in western Germany and France Rights and duties of married and non married couples in Germany Rights and duties of married and non married couples in France Employment status of western German and French couples with children in the year 2000 (per cent of couples in which one of the partners is employed) Summary of the events under study Relative risk of entering first union by type of union according to level and enrollment in education, western German women Transition to first union: Interaction between first school degree and current education, western German women Relative risk of entering first union by type of union according to women s activity status, western German women Relative risk of entering first union by type of union according to pregnancy motherhood status, western German women Relative risk of entering first union by cohabitation according to pregnancy motherhood status by birth cohort, western German women Relative risk of entering first union by direct marriage according to pregnancy motherhood status by birth cohort, western German women Relative risk of entering first union by type of union according to level of religiosity and parental family characteristics, western German women Relative risk of entering marriage after cohabitation according to woman s and partner s level and enrollment in education, western German women xiii

16 xiv LIST OF TABLES 7.9 Transition to subsequent marriage: Interaction between current level of education of the woman and partner s level of education at start of cohabitation, western German women Relative risk of entering subsequent marriage according to women s activity status, western German women Relative risk of entering subsequent marriage after cohabitation according to pregnancy motherhood status, western German women Relative risk of entering subsequent marriage after cohabitation according to pregnancy motherhood status by birth cohort, western German women Relative risk of entering subsequent marriage after cohabitation according to level of religiosity and parental family characteristics, western German women Unobserved heterogeneity: Standard deviation and correlation, western German women Relative risk of entering first union by type of union according to level and enrollment in education, French women Relative risk of entering first union by type of union according to women s activity status, French women Relative risk of entering first union by type of union according to pregnancy motherhood status, French women Relative risk of entering first union by cohabitation according to pregnancy motherhood status by birth cohort, French women Relative risk of entering first union by direct marriage according to pregnancy motherhood status by birth cohort, French women Relative risk of entering marriage after cohabitation according to woman s level and enrollment in education, French women Relative risk of entering subsequent marriage according to women s activity status,french women Relative risk of entering subsequent marriage after cohabitation according to pregnancy motherhood status, French women Relative risk of entering subsequent marriage after cohabitation according to pregnancy motherhood status by birth cohort, French women Unobserved heterogeneity: Standard deviation and correlation, French women Sample selection: western Germany Sample selection: France

17 LIST OF TABLES xv 9.3 Distribution of respondents according to the various levels of the time fixed covariates for western Germany. Absolute and relative number of respondents Distribution of respondents according to the various levels of the time fixed covariates for France. Absolute and relative number of respondents Distribution of time at risk (in person-months) according to the various time-varying covariates for western Germany. Absolute and relative number of person-months Distribution of time at risk (in person-months) according to the various time-varying covariates for France. Absolute and relative number of person-months Event history model for entry into first non marital cohabitation, western German women Event history model for entry into first non marital cohabitation, western German women, continued Event history model for entry into first direct marriage, western German women Event history model for entry into first direct marriage, western German women, continued Event history model for entry into first union, controlled for calendar time, western German women Event history model for entry into first union, controlled for calendar time, western German women, continued Joint model of entry into cohabitation vs. direct marriage as competing risks, western German women Event history model for entry into subsequent marriage after cohabitation, western German women Event history model for entry into subsequent marriage after cohabitation, western German women, continued Event history model for entry into subsequent marriage after cohabitation, western German women, continued Event history model for entry into subsequent marriage after cohabitation, controlled for calendar time, western German women Event history model for entry into subsequent marriage after cohabitation, controlled for calendar time, western German women, continued Event history model for entry into first non marital cohabitation, French women Event history model for entry into direct marriage, French women

18 xvi LIST OF TABLES 11.3 Event history model for entry into first union, controlled for calendar time, French women Joint model of entry into cohabitation vs. direct marriage as competing risks, French women Event history model for entry into subsequent marriage after cohabitation, French women Event history model for entry into subsequent marriage after cohabitation, controlled for calendar time, French women.. 275

19 Chapter 1 Introduction Recent changes in the pattern of union formation in Europe Over the last four decades, there have been remarkable changes in the pattern of union-formation and in the extent of having children outside marriage in many western European nations. Next to the spread of low fertility, there are hardly any other changes in family life which were as dramatic as the rapid rise in unmarried cohabitation 1 and non-marital births. Parallel, first marriage rates declined, the average age at marriage increased strongly and already existent marriages ended more often in divorces. The traditional sequence of a family formation pattern that has been characterized by early and wide-spread marriage and subsequent childbearing for about half a century, has become weaker or even disappeared. Unmarried cohabitation is not a new phenomenon in Europe. However, the reasons of forming such a relationship differ between periods. In a historic view, cohabiting unions have always been existent in most European cultural areas. Predominantly those were formed by poor people who could not afford to marry for economic reasons, by people who were opposed to marriage for ideological reasons or by people who chose to cohabit after a marital breakdown or the death of the former partner (Kiernan, 2001). In western Germany for example, after World War II older people often lived in cohabiting unions after a divorce or the death of their partner because they did not want to lose their alimony or widow s pension by remarriage 1 The terms unmarried cohabitation, cohabitation, cohabiting unions, consensual unions and non-marital partnerships are used synonymously. These partnerships consist of heterosexual partners who share a common household. 1

20 2 CHAPTER 1. INTRODUCTION (Nave-Herz, 2000). Since the 1970s, the group of people who cohabit is not restricted to particular subgroups anymore but is distributed among most social classes. Unmarried cohabitation has replaced marriage as the most frequent type of first union. While in some countries cohabitation mainly appears as a phase of transition which is restricted to young adulthood, in other regions of Europe also people in higher ages choose cohabitation as an alternative to marriage. Some authors argue that this development together with decreasing fertility rates and a postponement of childbearing can be regarded as a demographic shift, labelled as Second Demographic Transition (Lesthaeghe, 1992; van de Kaa, 1987). Others stress the importance of the growing economic independence of women and the associated reduced gains to marriage (Becker, 1981). In his extensive literature review on the development of cohabiting unions, Carmichael (1995) also mentions the availability of reliable contraception, prolonged educational enrollment, late 1960s youth radicalism, increasing secularization, increasing economic uncertainty of the young and the spread of individualistic values as some of the major explanations for the rise in consensual unions (Carmichael, 1995, p.61). To understand the reasons behind the increase in cohabiting unions and non-marital childbearing, cross-cultural comparisons can be very useful. There is still no clear convergence of family and union formation patterns in Europe (Kaufmann et al., 1997). Institutional factors, like the welfare regime of a society, specific policies and long-term cultural differences are primary sources for international differences in behavior. The overall framework in which an individual lives establishes a set of opportunities and constraints that shape the individual life course. This becomes finally visible as a regional pattern of union and family formation behavior. Therefore, it is reasonable to look at national family patterns that correlate with the social, cultural and institutional framework of a given society. The following work compares union dynamics in western Germany and France. The aim is to study the pattern of first union formation in both countries by describing and evaluating the most important factors and settings that determine possible differences between both countries. One of the main objectives will be a comparison of the social and institutional framework in France and western Germany and the search for explanations how these external conditions might affect union formation.

21 3 Why is it interesting to compare France and western Germany? Since the mid-1980s, a multitude of studies have been published which dealt with comparative welfare state research. Assuming that particular welfare regimes influence and structure life courses by shaping the educational and employment systems, many researchers compared the different systems in Europe and tried to detect systematic patterns. France and Germany have often been classified as conservative-corporatist welfare states (Esping- Andersen, 1990, 1999). Employment related and marriage related entitlements, the male breadwinner model as the predominant family model, the exclusion of non-employed, non-married women from social security and insufficient availability of public childcare are main characteristics of countries belonging to conservative welfare regimes. However, by analyzing tax systems, social security systems, child care arrangements and labor market behavior of women, international comparative studies found considerable differences within Esping-Andersen s classification (Anttonen and Sipilä, 1996; Gornick et al., 1997; Langan and Ostner, 1991; Lessenich and Ostner, 1995, 1998; Lewis, 1994; Orloff, 1993; Sainsbury, 1999). They mainly criticized his focus on the male perspective while neglecting the different degree of labor market integration of women. The grouping of the conservative welfare-state regime in particular has been criticized, as combining all countries that are neither liberal nor social-democratic. Especially concerning gender and family dimensions, the group of conservative welfare states presents itself very heterogeneous. Comparisons between France and Germany 2 add more information to the 2 Germany has a special history: during the four decades following the Second World War, Germany was separated into two countries: the Federal Republic of Germany (FRG - the western part of Germany) and the German Democratic Republic (GDR - the eastern part). East Germany had a state socialist system, a centrally planned economy, and socialist employment and family policies. West Germany, in contrast, had a multiparty parliament, a market economy, and a conservative corporatist welfare state (Rosenfeld et al., 2004). Both countries experienced an entirely different political, cultural, and demographic development before Germany was reunified in Even 20 years after the transition process, demographic patterns, attitudes and some institutional structures differ strongly between both parts of Germany. Therefore, we exclude the eastern German part from our theoretical and empirical discussion and analyze only the so-called Alten Bundesländer.

22 4 CHAPTER 1. INTRODUCTION different country-specific patterns that exist within this type of regime. Both countries differ immensely in terms of their family models, the extent of maternal labor force participation, the level of fertility, the share of non-marital births and the spread of cohabiting unions. German women live in consensual unions almost as often as do French women when they are young, but marriage is still much more dominant when they get older and particularly when children are involved. These so-called child-centered marriages (Matthias-Bleck, 2006; Nave-Herz, 1997) are explainable through a comprehensive system of historically grown family policies and institutional structures that hamper the economic independence of women. There are several institutional constraints and economic incentives in Germany that support the model of the married couple: co-insurance of married housewives by the health insurance of their husbands, higher tax reliefs for married couples in which one of the partners is not employed or works part-time (Ehegattensplitting), a parental leave scheme which until recently supported a longer exit from work after childbirth and weak father s rights in connection with illegitimate children before changes in legislation in Due to a low provision of child care facilities for children under six years of age and only half-day-schools for school age children, it is hard for western German women to combine their employment and their family life. Therefore they often quit or interrupt work for a few years to dedicate themselves to their family. The male breadwinner model with a non-working or part-time working mother is still very frequent. For western German women marriage does not only mean a personal commitment but it represents also an institution in which children can be raised and which provides financial coverage in case of separation. The interrelationship between marriage and parenthood seems to be very strong. In France we find family policies and an institutional framework that also support marital unions, but set a high value on the support of families with children as well. The number of children reduces the tax burden in French households. We find an almost complete assimilation of rights and duties for children, independently of the legal situation of their parents. France is nowadays an international leader in the provision of full-day preschools, allowing French mothers to be engaged in gainful employment and therefore be not dependent on their husbands income anymore. The higher share of French women working fulltime, the greater possibilities of getting child

23 5 care, the better acceptance of non-marital relationships in law and the earlier equalization of marital and non-marital births are factors that enable women to be independent from their husbands earning. The strong increase in nonmarital cohabitation in recent years, particularly in later life, as well as the high share of non-marital births indicate a changing paradigm - union status is not that important anymore, the interrelationship between marriage and childbearing seems to be much weaker than in western Germany. While cohabitation in western Germany has become a socially accepted, but only short-term prelude to marriage and is typically transformed into marriage when couples have a child (Blossfeld et al., 1999; Huinink, 1995), in France cohabitation has become an accepted alternative to marriage connected with a high rate of non-marital births (Leridon and Toulemon, 1995; Toulemon, 1997). The differences in demographic behavior and the apparent contrast in family policies provide a strong incentive to study France and western Germany in a comparative perspective. Though there are several studies which deal with a comparison of family and union formation patterns in France and Germany, most of them come from a political science or sociological perspective (for example Becker, 2000; Ehmann, 1999; Lessenich and Ostner, 1995; Reuter, 2002a; Schultheis, 1999; Veil, 1997, 2003), others from an economic perspective (Baclet et al., 2005). Until now, a demographic dimension which discusses family formation processes has been mostly neglected in a detailed country comparison, apart from some exceptions (Fagnani, 2002; Lauer and Weber, 2003; Onnen- Isemann, 2003). Particularly with regard to the rapid rise in unmarried cohabitations there are no extensive studies around. The present work closes this gap. Research objectives Most of the studies on marriage and cohabitation concentrate on women. Even though the underlying assumptions and theoretical implications are different, the fact that changes in union formation behavior are directly linked to the changing role of women is common to all theoretical approaches: the increase in female education, the growing labor-market participation of women and the consequential increasing options in life in the last decades. Many authors argue that the changing life concepts for women reduced the

24 6 CHAPTER 1. INTRODUCTION desirability to marry since gender-specific division of labor became less important and opportunity costs of childbearing increased drastically (Becker, 1981). As a consequence, marriage rates decreased and divorce rates increased. Others emphasize the fact that women s greater economic independence not necessarily leads to a decline in the proportion of women ever married but mainly to an increase in delayed marriage. They attribute delays in marriage and the rising age at marriage to the improved bargaining position of women. Greater economic independence allows women to search longer for an acceptable match and reduces the probability to remain in unhappy marriages (Oppenheimer, 1988). Well-educated women with substantial earning power can incorporate premarital cohabitation into search and bargaining processes because cohabitation provides good opportunities to observe men s earnings potential and willingness to share household and childrearing tasks before a more binding relationship such as marriage (Cherlin, 2000). Next to the increase in women s education, the prolongation of education is also discussed as one factor: It is not the rise in human capital investments of women that leads to delayed marriages but women s longer participation in the educational system (Blossfeld and Huinink, 1991; Blossfeld and Jaenichen, 1992). Apart from the changing role of women over the last decades, men s situation, especially their education and economic position, should not be underestimated. Men s earning potential and their career opportunities play an important role for the timing of marriage of both men and women as well. Historically, women s marriage timing has primarily been a function of young men s economic characteristics and not the other way round (Oppenheimer, 1997). Additionally, in times of economic uncertainties it is often necessary to have more than one earner in the household or pool earnings (Oppenheimer, 1997). However, to include men into the analysis would go beyond the scope of this work, therefore men s union formation behavior is not discussed in detail but included to some extend via the women s partner s characteristics. The major objective of this work is to analyze the two most discussed factors that are made responsible for changes in union formation behavior: the impact of women s education and the impact of childbearing on union formation behavior. Since a rise in education and female employment per

25 7 se does not necessarily lead to a growth in women s economic independence, we also consider and discuss country specific structures which might hinder women and in particular mothers to translate their improved educational opportunities into an increase in their labor force attachment. Different welfare state regimes and their embedded family policies produce different patterns of labor market integration of both men and women and strengthen or weaken therewith women s dependency from the institution of marriage. They also influence legislation on marriage and cohabitation which again impacts the individual decision regarding union formation. Therefore, we are strongly interested in the effect of different family policies and institutional structures on union formation behavior. In this study, we exploit variation by country and over time to study the impact of family policies. Certain hypotheses on the effect of the contextual framework on marriage and cohabitation have therefore been developed. In western Germany, women have only limited access to employment and marriage presents the main institution of economic protection for women with children. Under these conditions we assume that unmarried parenthood will be avoided and a binding and legal confirmed institution like marriage will be preferred. In France, where female employment is encouraged and women are supposed to keep their living without being dependent on a family member that acts a the main breadwinner, an institution like marriage is not a necessary precondition anymore - not even when children are involved. It might more and more be replaced by cohabitation - at least in the beginning. In the empirical part of this work, we investigate possible changes in the effects of certain covariates such as education, employment or childbearing over time by basing our analytical framework on the life course perspective (Giele and Elder, 1985; Kohli, 1985; Mayer and Schoepflin, 1989). We use the technique of event history analysis to analyze individuals, their positions and status changes over a particular length of time. A competing-risk framework is used to study direct marriage versus non-marital cohabitation. For cohabites, a model on the determinants of subsequent marriage formation is estimated. How does the occurrence of first cohabitation and/or first marriage is effected by other events such as childbearing, educational attainment or employment? Have these effects changed over the last decades? Are cohabiting unions a prelude or an alternative to marriage? In addition, we

26 8 CHAPTER 1. INTRODUCTION analyze the extent to which the conception of a first child determines subsequent marriage rates in both countries. The effect of childbearing on union formation behavior can only be displayed by a certain extend since only actual behavior can be analyzed within our models. Whether couples marry because they plan to get children cannot be measured empirical. However, we are able to detect whether the interrelationship between union formation and pregnancy/birth is stronger in one country than in the other. By analyzing both events as interrelated processes we try to find out whether observed and unobserved individual characteristics simultaneously influence first birth and first marriage. Cross sectional data cannot answer these questions. By using life course analysis, we can reconstruct the biographic embedding of marriage or cohabitation into the individual life course. To be married or to live in a cohabiting union does not mean the same in France and western Germany. The following work is going to shed some lights on these differences by 1. closely analyzing the contextual impacts on union formation, 2. looking at the timing of union formation in the life course and 3. interpreting key factors, particularly the effect of female education and employment and the incidence of a pregnancy, and their influence on the decision to marry or not to marry. We use two surveys with comparable longitudinal information collected trough retrospective accounts of life histories: the German Familiensurvey (Familysurvey) conducted in the year 2000 and the French Etude de l Histoire Familiale (Study on Family History) conducted in The period of time covered in the empirical study stretches from the 1960s until the end of the 20th century. Outline of the study Chapter 2 introduces general demographic indicators and an overview on marriage and childbearing pattern in western Germany and France. Chapter 3 presents literature on the determinants of marriage and cohabitation and discusses possible explanations for recent changes in union formation

27 9 behavior. Chapter 4 gives a detailed overview on social and structural conditions in the two countries which will help to understand why people might prefer a particular type of union or not. On the basis of the theoretical and institutional background, main research hypotheses will be presented (Chapter 5) that lead to the empirical part of this work. This part starts in Chapter 6 which introduces methods and data sets and discusses sample selection, definitions of events as well as selection of covariates and previous research. This is followed by the empirical part of this work in Chapter 7. Main descriptive results are presented in Section 7.2 in which the transitions into different kinds of unions by age and cohort are analyzed. Cumulated incidence curves and the multivariate analyzes of entry into first union are presented in Section 7.3 for western Germany and Section 7.4 for France. In Section 7.5 we summarize the empirical results. Chapter 8 combines theoretical considerations with our results and concludes with an outlook to future research challenges.

28 10 CHAPTER 1. INTRODUCTION

29 Chapter 2 Demographic developments Germany and France are neighboring countries situated in Middle/West Europe. Both countries share some common features in the process of adolescence and within the family formation process. The entry into sexual life starts within the same age in France and Germany, men and women are between 17 and 18 years old in both countries (Bozon, 2003). Also, the use of contraception is widely spread: three-quarter of all women use contraceptives (United Nations, 2003). People in both countries need longer to become economical independent from their parental home than did the generations before them. They remain longer in education: the median age at leaving school in France and Germany increased over cohorts. In France, the age at leaving school increased from 14 years for people born before 1930 to 20 years for those born around the year 1970 (Robert-Bobée and Mazuy, 2003). Also in Germany, the median age at leaving school and starting to work has increased (Konietzka and Huinink, 2003). Table 2.1 displays a summary of some of the most important demographic indicators concerning marriage and family and its changes during the last decades. 11

30 12 CHAPTER 2. DEMOGRAPHIC DEVELOPMENTS Table 2.1: Development of selected demographic indicators in western Germany and France W-Germany France total first marriage rate (per 100 women) total divorce rate (per 100 marriages) total fertility rate (per woman) mean age at female first marriage (in years) mean age at birth of first child (in years) proportion of non marital births (in %) proportion of ever married women by age 50 ( in %) , 2 marriages <=25 years, , , 5 birth cohorts 1930 and 1965 Sources: Desplanques (2008); Pla and Beaumel (2010); Prioux and Mazuy (2009); Statistisches Bundesamt (2010), some data for western Germany has been provided by personal contact from the Statistical Office of Germany Parallel to the increase in ages at leaving education and starting working life, the mean age at becoming pregnant 3 or getting married for the first time has risen remarkably in both countries. In both countries, women who married in the year 2008 are on average around six years older than 48 years ago. Also the mean age at first birth increased up to four or five years. Total first marriage rates almost halved in both countries. Unions also tend to be shorter lived, as separation and divorce become more common. In both countries, divorce rates increased. As marriages became more seldom, the proportion of unions that began as non-marital partnerships and the number of births outside of marriage increased, particularly in the last twenty years. Starting from the same level in 1960, nowadays France exhibits a much greater proportion of non marital births than western Germany: 51.6 % of all children born in the year 2008 have been born to non married parents in France compared to 25.8% in western Germany. The proportion of women who have ever been married by age 50 decreased as well, however, still almost two-thirds of all women born in 1965 married at least once in their life in 3 Since numbers for western Germany have only been available for births of children born within marriage, the mean age at birth of first child is overestimated. Estimations independent of marital status made by Kreyenfeld (2002a) for earlier years showed a mean age at first birth of 27.1 for the year 1995 instead of the official number of 28.1.

31 13 both countries. We display two standardized indicators in more detail to show the decrease and postponement of marriage over time in both countries. In figure 2.1 we see crude marriage rates by age and calendar year. Figure 2.1: Crude marriage rate, western Germany France Source: Council of Europe (2001) Crude marriage rates express the number of marriages formed each year as a ratio to 1000 people, expressed with respect to population size. This measure naturally disregards families based on informal partnerships and other types of legal unions as well as married but separated spouses. Crude marriage rates are also strongly affected by the demographics of the population in question. Both countries display a strong decrease in marriage rates after the 1960s. Until the end of the 1970s marriage was more pronounced in France than in western Germany. However, this changed in the early 1980s when French marriage rates dropped well below the level of western Germany. Only in recent times both countries seem to converge. Total female first marriage rates are less affected than crude marriage rates by the overall demographics of the population. The total female first marriage rate (TFFMR) is estimated as the sum of age-specific marriage rates of single women aged It shows how many percentage of single women would marry if current marriage conditions remained. In times of sinking marriage ages and catch-up-marriages (as was the case in the 1950s),

32 14 CHAPTER 2. DEMOGRAPHIC DEVELOPMENTS this synthetic index can reach values above 100. In France and western Germany, we observe an almost steady decline in female first marriages since the end of the 1960s (figure 2.2). Figure 2.2: Total female first marriage rate, western Germany France Source: Council of Europe (2001) Before that time, marriage was universal almost every woman got married during the 1950s and 1960s. After the sharp decline during the 1970s, in western Germany the TFFMR levelled off during the 1980s and remained since then on a level of 60% until the year In France, the decline in female first marriages started earlier than in western Germany but exhibited higher numbers during the 1970s. After the beginning of the 1980s the TFFMR declined and dropped below the level in western Germany. Starting in 1995 we can observe an upswing in the TFFMR. This increase has been identified as a timing effect caused by a change in tax law which encouraged couples with children to legalize their union and thereby pay less tax. As a consequence, marriage rates of cohabiting parents increased strongly between 1996 and 1997 (INSEE, 2002, p.13). It was, however, only a short term increase with another peak in 2000 caused by the millennium effect. In last years, the TFFMR has been decreasing again: Since the 2000 peak of 60%, the index has been gradually falling to 51% in the year 2007 (Prioux, 2008). However, also the TFFMR must be interpreted carefully because it gets strongly influenced by changes in the age at marriage or by period